Volume Rendering of AMR Simulations

High-resolution simulations often rely on the Adaptive Mesh Resolution (AMR) technique to optimize memory consumption versus attainable precision. While this technique allows for dramatic improvements in terms of computing performance, the analysis and visualization of its data outputs remain challenging. The lack of effective volume renderers for the octree-based AMR used by the RAMSES simulation program has led to the development of the solutions presented in this paper. Two custom algorithms are discussed, based on the splatting and the ray-casting techniques. Their usage is illustrated in the context of the visualization of a high-resolution, 6000-processor simulation of a Milky Way-like galaxy. Performance obtained in terms of memory management and parallelism speedup are presented.